Page 3 of 4 FirstFirst 1 2 3 4 LastLast
Results 41 to 60 of 70

Thread: GigaBump an MDD

  1. #41
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    Now about temperatures inside. Since the Giga upgrade does not have an onboard CPU temperature sensor, I can only use Hardware Monitor drive temperatures for drives connected to the mobo ATA66 and ATA100 buses. The two PATA drives connected to the Acard 6885M do not appear in Hardware Monitor, as expected. Additionally, I have a Radio Shack digital indoor/outdoor thermometer with the outdoor probe mounted between the vertical ATA100 drive cage and the Giga CPU fan.

    Fully warmed up for the few days of constant use typically, shutdown only late night:
    ATA66 cage drive 37C/99F (Hardware Monitor)
    ATA100 cage drive 31C/88F (Hardware Monitor)
    RS probe inside MDD 35C/95F
    RS room temperature 25C/77F
    k

  2. #42
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    I have long suspected that a CRT, any old CRT, would run Xbench faster than an LCD, I suspect because of response time of an LCD, even the 14ms of my 23-inch Cinema HD Display. Fast LCD response times are now down to less than 5ms for some monitors. CRT response time is zero.

    Anyway after trying my LCD in Xbench 1.2, I tried again with a CRT, very old ViewSonic GS790-2. All GigaMDD runs 10.4.5:
    v1.2 LCD best 73.34, range 71.65 - 73.34
    v1.2 CRT best 75.60. range 74.27 - 75.60

    v1.1.3 LCD previous best 228.43
    v1.1.3 CRT best 229.36, range 228.25 - 229.36 just run twice

    The two best scores for those versions of Xbench will be posted to Xbench later. Going to try the CRT on my G5 later. k

  3. #43
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    Results for Xbench v1.2 CRT 75.60 here
    v1.1.3 CRT 229.36 here k

  4. #44
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    I have been pounding on this GigaMDD for several days now, mostly testing SATA drives using a SeriTek/1VE4 and trying to find the sweet spot for SATA RAID on this 167MHz bus. But that is for another post.

    When installing the SeriTek, I checked my other PCI cards:

    Slot-1 - (AGP) ATI 9800 Pro ME (128MB)
    Slot-2 - Acard 6885ME 4-channel running 64-bit/33MHz
    Slot-3 - M-Audio Revolution 7.1 32-bit/33MHz
    Slot-4 - SeriTek/1VE4 running 64-bit/33MHz
    Slot-5 - 7-Port NitroAV FW800/USB 2 Combo Card running 64-bit/33MHz
    external, 2 FW800 ports and 3 USB 2 ports
    internal, 1 FW400 port and 1 USB 2 port

    Also removed all 4 512MB memory sticks and installed 4 Gurus MDT DDR 1024MB PC400 CL 2.5 sticks. Yeah I know the memory controller can't see more than 2048MB but I had to try and along the way I confirmed something else again.

    Nothing, apps nor utilities saw more than 2048MB. Tiger System Profiler reported:
    J21 - 1024MB
    J22 - 1024MB
    J23 - empty
    J20 - empty

    So the memory controller will only look for 2048MB for sure. But secondly, it looks at J21 and J22 before J23 and J20. Way to go Apple and why is J20 looked at after J21 and J22? I've asked this before. k

  5. #45
    Join Date
    Aug 2003
    Location
    Under the Midnight Sun
    Posts
    656

    Default 64-bit Seritek

    Kaye, I am curious how this is found out? "running 64-bit/33MHz" Does the profiler tell us this info?

    I have a MDD 1.25, as well as a 1se2 and 1Ve4, but was unaware that the later would offer 64-bit support (just found that info on the Firmtek site)...

    How do we know they are really in 64-bit mode?

    Rob
    MacBook Pro Uni 2.4GHz with express card slot

    iMac G5

  6. #46
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    Rob,

    No the profiler does not tell you. This helps me. I have printed that page. Also, as you say, FirmTek and Acard say so as does the box for that Acard model. Of course we know that the MDD PCI runs 33MHz even if we have a 66MHz or higher (100MHz or 133MHz) card.

    On my G5 I am very careful to only have cards that are 64-bit 100/133MHz only. In my G4, the only card that is 32-bit/33MHz is the Revo 7.1. It is a very short card. But it does not affect the bus that I can tell. I suppose if I were listening to iTunes while running a SATA test or the Pshop test, I might see a performance loss, but as ataMan has said in the past, you never know what if any loss might occur. Each card has to be looked at individually, all 32-bit/33MHz cards are not the same.

    Another point ataMan said about the G4 was not to run two SeriTek/1S2 cards and RAID together the drives on both cards, somewhat degraded I/O particularly early G4s. I found that to be true on the MDD. I ran another test on the G4 with separate RAIDs running simultaneously (QB2.1 extended test), one RAID on a UL3D-33, the other on a SeriTek/1S2, and you had to be fast to get the tests going near simultaneously with two QB2.1 apps. Both RAIDs slowed down as soon as both were running the tests. Once the tests were complete, the UL3D recovered immediately for another single test while the SeriTek would not recover and required a reboot.

    Best way I know of regardless of what the box or website says is see if what you are testing performs as it should. Another card could be the cause also. k

  7. #47
    Join Date
    Aug 2003
    Location
    Under the Midnight Sun
    Posts
    656

    Default OKy Doky

    Ich verstehe, danke
    MacBook Pro Uni 2.4GHz with express card slot

    iMac G5

  8. #48
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    Trying to find the sweet spot for the GigaMDD with my SATA drives, 4x Raptor74s and 2x Raptor1500s while considering the 167MHz bus and other PCI cards. All tests with SeriTek/1VE4 and external cabling. 10.4.5 and SoftRAID 3.3. Where RAID, 128KB SU. Results rounded for read/write. G5 are previous tests on the DB with same SeriTek/1VE4.

    QuickBench» 2.1 Test Results File
    ©2000-2005 Intech Software Corp.
    Test file created on (various dates and times)
    Test Volume name: (various)
    1x Raptor74
    100MB - 64/74
    Same as G5

    2x Raptor74
    100MB - 120/157
    Same as G5

    1x Raptor1500
    100MB - 73/94
    G5 - 73/92

    2x Raptor1500
    100MB - 151/180
    G5 - 128/218

    4x Raptor74
    100MB - 206/185
    G5 - 231/341

    1x Raptor74/2x Raptor1500 (this one seems to be the sweet spot)
    100MB - 205/185
    G5 - no results on the DB

    2x Raptor74/2x Raptor1500
    100MB - 201/185
    G5 - no results on the DB

    I am going to replace the Revolution 7.1 card with a UL3D and test the 4x Atlas 36GB 15K II's. But that is for another day or week. k

  9. #49
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    I had been working on the GigaMDD for several days interrupted by a trip to the doctor Monday for a painful and sudden growth on my back. He immediately had me admitted to the hospital for surgery to remove a deep sebaceous cyst, so deep and infected that the surgeon is not sure he got it all. It grew in about two days during the weekend. Now back in action.

    Before this started I removed the unrecognized 2x 1GB sticks of RAM leaving the remaining 2x 1GB sticks of Gurus MDT DDR 1024MB PC400 CL 2.5 sticks in J21 and J22. Did an Open Firmware reset.

    Also did the update from 10.4.5 to 10.4.6. Disk Utility repaired permissions and verified the drive which is one volume. Rebooted and shortly thereafter everything ground to a halt. Just after the desktop was complete, got a kernal panic and the internal fans went to super high speed. Shutdown with the front button and walked away thinking possibly another mobo failure. Thought about it for awhile and decided to see if I could startup with another boot drive using the option key.

    First picked a backup drive with 10.3.8 (and 9.2.2). That worked fine. Then rebooted to another boot drive with 10.4.5 (and 9.2.2). That also worked fine. So I started up my FW400 drive and used DiskWarrior 3.0.3 off of it to repair the 10.4.6 Tiger only drive. It reported a problem with the Desktop and fixed it. No problems with the 10.4.6 Tiger only drive after that. Thankgoodness for DiskWarrior. Never seen that before. Five volumes, two folders, and an alias to another folder on the Tiger only drive Desktop.

    Also removed the M-Audio Revolution 7.1 32-bit/33MHz PCI card from Slot-3. I thought long and hard about which UL3D I wanted to replace it with, UL3D-33MHz or 66MHz. Decided on the 66MHz version which would run 64-bit/33MHz in the MDD. This for a test later of my 4x Atlas 15K II Burly later. Did an Open Firmware reset and went to ATTO for all the latest installers, driver, Config Tool, and flash file for the UL3D-066.

    So now my PCI cards:
    Slot-1 - (AGP) ATI 9800 Pro ME (128MB)
    Slot-2 - Acard 6885ME 4-channel running 64-bit/33MHz
    Slot-3 - ATTO UL3D-66MHz 64-bit/33MHz
    Slot-4 - SeriTek/1VE4 running 64-bit/33MHz
    Slot-5 - 7-Port NitroAV FW800/USB 2 Combo Card running 64-bit/33MHz
    NitroAV external, 2 FW800 ports and 3 USB 2 ports and internal, 1 FW400 port and 1 USB 2 port

    First a test of the revised Cinebench 9.5 which I'll call rev1. Comparing the two 9.5 versions:

    CINEBENCH 9.5//9.5rev1
    **************************************************
    Tester : kyum//same
    Processor : GigaMDD G4-1.42GHz//same
    MHz : 1420//same
    Number of CPUs : 2/same
    Operating System : 10.4.5//10.4.6

    Graphics Card : Radeon 9800ME 128MB//same
    Resolution : 1920x1200//same
    Color Depth : Millions//same
    **************************************************
    Rendering (Single CPU): 159 CB-CPU//165
    Rendering (Multiple CPU): 255 CB-CPU//297

    Multiprocessor Speedup: 1.60//1.80

    Shading (CINEMA 4D) : 162 CB-GFX//168
    Shading (OpenGL Software Lighting) : 390 CB-GFX//501
    Shading (OpenGL Hardware Lighting) : 797 CB-GFX//1158

    OpenGL Speedup: 4.92//6.89
    **************************************************
    Speedup occurs in all tests. Have completed testing QB2.1 and ZoneBench for 2x and 4x Raptor 1500 drives for scratch but have to put that together. k

  10. #50
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    SeriTek/1VE4, SoftRAID 3.3, where RAID, 128KB SU, 10.4.5 (previous tests), 10.4.6 for 3x and 4x Raptor1500 tests.

    QuickBench 2.1 Test Results File
    ©2000-2005 Intech Software Corp.
    Test file created on (various dates and times)
    Test Volume name: (various)

    1x Raptor1500
    100MB - 73/94 (previous test)
    G5 - 73/92

    2x Raptor1500
    100MB - 151/180 (previous test)
    G5 - 128/218

    3x Raptor1500 (418GB)
    100MB - 210/185
    G5 - 187/346

    4x Raptor1500 (559GB)
    100MB - 212/186
    G5 - 246/456



    k

  11. #51
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    UL3D-066, SoftRAID 3.3, 128KB SU, 10.4.6.
    4x15K II 137GB
    QuickBenchÈ 2.1 Test Results File
    ©2000-2005 Intech Software Corp.
    Test file created on Wednesday, April 26, 2006 at 9:03:55 PM
    Test Volume name: 4x15K2sr
    4x Atlas 15K II
    100MB - 200/200
    G5 - 328/401




    Also striped 4xRaptor1500 and 4x15K II 274GB, SeriTek/1VE4 and UL3D-066, SoftRAID 3.3, 128KB SU, 10.4.6.
    8xDrives
    100MB - 223/181
    G5 - 535/679



    k
    Last edited by kaye; 04-27-2006 at 10:26 PM.

  12. #52
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    Comparing the Radeon 9800ME 128MB with the 256MB version. I purchased the card while there was an ATI mail in rebate mainly for X-Plane. Cranked up the bells and whistles to near maximum (Xplane will reduce those settings to maintain at least 20fps). With the 128MB card on approach the sim would stop then jerk forward seemingly several hundred feet to catch up and did this throughout the approach. With the 256MB card at the same settings (Xplane still reduces those settings but to 23fps this time) but the approach was totally smooth and more eye candy. I am pleased with that.

    The 128MB card is identified as ATY, R350. So is the 256MB card. ATIccelerator II identifies the 128MB card:
    Processor 378.00 MHz
    Memory 337.50 MHz
    The 256MB card shows the exact same speeds so the GPU and the memory are the same (tho the memory is a different brand), just 2x the memory. Slight increase in Shading with Cinebench:

    CINEBENCH 9.5rev1//9.5rev1
    **************************************************
    Tester : kyum//same
    Processor : GigaMDD G4-1.42GHz//same
    MHz : 1420//same
    Number of CPUs : 2/same
    Operating System : 10.4.6//same

    Graphics Card : Radeon 9800ME 128MB//256MB
    Resolution : 1920x1200//same
    Color Depth : Millions//same
    **************************************************
    Rendering (Single CPU): 165 CB-CPU//165
    Rendering (Multiple CPU): 297 CB-CPU//297

    Multiprocessor Speedup: 1.80//1.80

    Shading (CINEMA 4D) : 168 CB-GFX//168
    Shading (OpenGL Software Lighting) : 501 CB-GFX//503
    Shading (OpenGL Hardware Lighting) : 1158 CB-GFX//1165

    OpenGL Speedup: 6.89//6.91
    **************************************************
    k

  13. #53
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    GigaMDD AltiVec Fractal Carbon and Power Fractal comparison tests
    Graphics Cards : Radeon 9800ME 128MB//256MB VRAM
    **************************************************
    Altivec Fractal Carbon v1.2
    128MB // 256MB
    1.2s, 11110.4MF (10.4.5)//1.2s, 11264.7MF (10.4.6)
    **************************************************
    Power Fractal v1.4
    256MB
    1.2s, 11264.7MF (10.4.6)
    -----------
    These tests and results do not behave the same way as on my G5 where AFC yields faster results than PF because of a slight delay in starting PF. Go figure. Perhaps the G4 is slower to start AFC and PF equally. k

  14. #54
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    Took me 2 days to get this Giga MDD current. Ran 3 Xbench 1.3 tests. Cinebench to follow. k

  15. #55
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    No change to Power Fractal results with the Radeon 9800 ME 256MB card.

    CINEBENCH 9.6
    ************************************************** **

    Tester : kyum

    Processor : GigaMDD1.42GHz DP
    MHz : 1420MHz
    Number of CPUs : 2
    Operating System : 10.4.8

    Graphics Card : Radeon 9800Pro ME-256MB
    Resolution : 1920x1200
    Color Depth : Millions

    ************************************************** **

    Rendering (Single CPU): 164 CB-CPU
    Rendering (Multiple CPU): 295 CB-CPU

    Multiprocessor Speedup: 1.80

    Shading (CINEMA 4D) : 169 CB-GFX
    Shading (OpenGL Software Lighting) : 506 CB-GFX
    Shading (OpenGL Hardware Lighting) : 1148 CB-GFX

    OpenGL Speedup: 6.81

    ************************************************** **
    k

  16. #56
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    External 4xRaptor150, 558GB, 4x channels, SeriTek/1VE4 PCI card, SoftRAID 3.6, 32KB SU. Not the fastest for QB nor ZB but optimized for Photoshop CS2.



    k

  17. #57
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    I updated to 10.4.10 and ran Geekbench. Score 1190. http://browse.geekbench.ca/geekbench2/view/11056

    Not the fastest MDD on the block but near the top. Some faster with a different processor or overclocked beyond 1.42GHz or CL2 memory or just faster, a variety of reasons. Anyway, I'm happy with the result. k

  18. #58
    Join Date
    Mar 2003
    Posts
    203

    Default Is GigaDesigns out of business?

    I have been unsuccessful in trying to contact Tim at GigaDesigns for several months now, both by email and (voice) telephone. There is still a recording on the answering machine, but no one has returned my phone calls or emails.

    It looks like Tim has dropped off the face of the earth.

    ???

  19. #59
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    I don't know their status. They still have a website and support info:
    http://www.gigadesigns.com/index.asp...on=Custom&ID=6

    but you probably have all that. Maybe Rick knows what happened. I thought they were purchased by OWC. k

  20. #60
    Join Date
    Mar 2003
    Posts
    203

    Default

    Thanks, Kaye.

    I guess I will call OWC tomorrow and ask if they know anything.

    Cheers

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •